Promoting fire safety in innovating design

of electric vehicles: the example of the

EU-funded DEMOBASE project

G. Marlair1, A. Lecocq1, B. Truchot 1 P. Perlo2,

M. Petit3, D. N’Guyen4 & P. Desprez4

1INERIS, France; 2IFEVS, Italy, 3IFPEN, France, 4SAFT, France

5th Int. Conf. On Fires in Vehicles – FIVE 2108, Borås, Sweden, October 3-4, 2018

❑ Project ID & Rationale of DEMOBASE

❑ Fire research dedicated strategy

and relating commitments in the project

❑ Results achived at 1/3 project time scale

▪ Focus on analytical approaches

▪ Status of experimental and modeling approaches

❑ Conclusions/ perspectives

Outline

❑Real market penetration of e-mobility worldwide

depending on:▪ drastic reduction of costs

▪ increased performances not forgetting safety

▪ increased availability

faster evaluation and integration of innovative technology of key

components of Evs

❑H2020 DEMOBASE project▪ DEsign amd MOdeling for Improved Battery Safety and Efficiency

▪ stands as a closed loop project response in direction of market driven demand

▪ Objective:

• implementation of innovative and continous process for integration of new active

materials, component and cells iin EVS by use of multi-scale modeling and

testing integrating battery management, and taking due account of ageing, safety

and recycling issues and challenging built car prototype against KPIs

DEMOBASE Project ID and Rationale

❑ UE grant # 769900, RIA type of project

❑ Priority addressed: GV7-H2020: multilevel modelling and testing

of electric vehicles and their components

❑ Consortium: 11 partners

❑ Scientific coordination SAFT

❑ Project management: K&S GmbH Projektmanagement

❑ WP leader safety: INERIS

❑ Estimated project cost: 7,451,520 €

❑ duration: 36 months, started 1st Oct. 2017

❑ URL: www.demobase-project-eu

❑ Contact: info@demobase-project.eu

DEMOBASE EU-funded project data

http://www.demobase-project-eu/

Overall organization of the project and relation to safety

➢ Need for better understanding and

modeling of the TR for each chemistry,

including most reactive cathode materials !

➢ Better battery SOH prediction and improved

safety management devices

➢ System safety also needed, BMS related, but not only

The fire safety challenge in relation of the thermal

runaway hazard

Feng et al, Ener. Storage Mater.10 (2018), 246-267

❑Preliminary paperwork (first round completed):

▪ Accident review

▪ PRA

▪ BMS (fire) safety management optimisation (pending)

❑Experimental approaches on key EV components▪ Cell components, battery cells & modules, packaging materials

▪ Feeding component/cell selection, package design and modeling needs

▪ Started in 2018

❑Fire safety related Modeling

▪ Multiphysics, multi-scale, multi-tools approaches, multi-objective

▪ Pluri-partners synergism seeked for

▪ Work started at IF-PEN, SAFT and INERIS on ES level

Fire safety research dedicated strategy in the project

❑Methodology based of previous studies▪ Review of past accidentology of lithium-ion batteries with a focus on Evs

▪ Scenario-based PRA, starting from EV/energy storage design to end use

and recycling

Risk Assessment of Lithium-ion Based Energy Storage Systems

Preliminary Risk Analysis (EV deployment)From a full value chain perspective

❑ Examination of incidents/accidents from lithium batteries (LI-ion

and Li Metal Polymer) identified from various sources relating to

e-mobility

▪ Existing official databases (eg ARIA), or originating from blogs

(www.wreckedexotics.com/ )

▪ Expert network

▪ Internet searches with web browsers

❑ Possible biais in the analysis due to uneven access to info and

reliability issues, at world level

❑ Databases built up in the project did not dig in issues about

consumer market batteries, …)

Accidentology review: methods

http://www.wreckedexotics.com/

Accidentology: analysis performed vs location in the value chain

Example: partial view of incident logs regarding EVs or hybrid buses during use phase

➔Similar tables set up for all stages of the EV value chain (design,manufacture, transport, storage, use, recycling)

❑ Some incidents involving Li-ion batteries still arising from mishaps at design/quality

control stages (Innovation ➔ risks)

❑ Thermal and mechanical protection, also positioning of battery important to avoid

abuse conditions leading to accidents

▪ (nearby fire, car crash of runover, impact with sharp objects on roadways…)

❑ electrical protection against water/moisture driven short circuit importance also

revealed from EV accidentology

❑ Accidentology also reveal the importance of the alert function in case of EV battery

failure to allow fast abd safe evacuation

❑ Fire risk management in recycling sites may lead to significant damage in case of

fire due to projections of battery components and release of fire brands and toxic

smoke

❑ Fire-fighting of battery fire may be very difficult and require training, late reignition

of EV batteries after incidents must be anticipated

❑ no sign in our view of increased fire hazard as compared to ICE cars, but fire

hazard typology a bit different and fire prevention/protection need to be customized

to EV pecularities

Accidentology review: main observations

❑ Collecting, and sortingTesla cars

crash/runover and relating fire

events according to incident

outcomes have been performed

❑ Analysis in terms of

▪ circumstances

▪ Seriousness

▪ Comparison with ICE cars

(tentative)

❑ As for all other EVs, Tesla accidents

do not necessarily end up by a fire

event!

Fire Accidentology review on EVs: a focus on the Tesla models

❑ Models investigated: Tesla Roadster (Lotus platform), Model S, (sedan) Model

X (SUV), Model 3

❑ 21 reported fires, most info from web sites and Tesla media reports + one

scientific report (about 3 first scenarios in USA and Mexico)

▪ need to be related to some 300,000 -350,000 Tesla cars sold so far in some 5 years

▪ some 150,000 car fires in USA, some 30,000 car fires in France on a annual basis

▪ 131 Tesla car crashes/overturn reported (www.wreckedexotics.com/ )

• Model S: 79 ; Model X: 16 ; model 3: 3; model Roadster: 34

▪ 100% of fire deaths in Tesla fires are relating to post crash fires, currently

❑ To be noted: early recall of over 400 Roadster model 2010 carsl for fire

hazard reasons pertaining to inadequate battery cable routing

Tesla post-crash fire events/vs all Tesla

post-crashes (with/without fire

event )trend: ~ 12,5%

comparison with ICE car fire trends

from NFPA 2010 stats:

- post crash fires: 3% of all vehicle firesholding for 58% of vehicle fire deaths

Tesla EV fires: some trends and statistics

10; 48%

3; 14%

5; 24%

3; 14%

distribution of 21 Tesla fires according to main cause

post crash/turnover fires

spontaneous fire (driving)

spontaneous fire (park)

fires due to metal partimpact

http://www.wreckedexotics.com/

❑ Re-rating previously identidied scenarii seriousness making use of a risk

matrix from inital 2011 study (revised already in 2013)

❑ Completing analysis in terms of new pertinent incident scenario deserving

examination at light of field experiene

❑ Back-up info to all stakeholders of

DEMOBASE, according to involvment

in concerned EV building-blocks for

due consideration

PRA update from original INERIS study: scope, methods and follow-

up

Revised PRA ➔ renumbering and re-rating incident scenarios through

establishment of relating database (abstract of the 52 scenarii identified)

PRA EV update: major results

2013

❑ 21 scenarii found less critical thaninitial PRA 2011 study,

❑ 3 scenarii found more criticalthan in initial INERIS PRA study,

❑ 26 unchanged criticity rating

❑ 2 new scenarii

2018

❑Priorities:▪ Contribute to qualify electrode materials/cell options in terms of

fire safety aspects (thermal and electrical abuse)

▪ Provide calibration data for multi-physics TR prediction model,

cell level on fresh and aged cells

▪ Provide calibration data for CFD computation (cell and module

level)

▪ Test reaction-to-fire of key battery pack casing/insulation

material for EV battery integrator

Fire / thermal abuse testing strategy

Testing approaches Protocol and testing device

❑ Testing device and instrumentation

Thermal tests performed in BTC 500

from HEL

• 2 thermocouples for regulation measurement positionned on each side of the cell• 4 thermocouples around the cell• 4 others thermocouples inside the equipement• Cell voltage measurement

+

-

1

2

Cordon

6

3

5

4

• Heater wire enrolled around the cell• Cell positionned at the center of the

equipment on a support• Cell charged at 100% before test

❑ Use of FPA apparatus (ISO 12136) coupled with FTIR

instrument (18 gas exploitation method):

▪ tests carried out on pack insulation material candidate

Testing reaction to fire on key combustible pack materials

❑ Protocol Example of result from early testingon fresh cell in DEMOBASE

Protocol and testing device

250°C

Starting at the operating temperature of the sample, the temperature shall be increased in 5°C increments (at a ramp of 5°C/min) and held at each step for a minimum of 30 minutes, or until any self-heating is detected. If self-heating is detected, the chamber temperature shall track the cell temperature until the exotherm becomes stable. The temperature is then increased to the next 5°C increment and continued as described above until (1) additional self-heating is detected, (2) the temperature reaches 200°C above the operating temperature of the sample (250°C), or (3) a catastrophic event occurs.

« Heat-Wait and Search program »

▪ Adiabatic conditions, heat-wait-search process (ARC) to characterize on-set temperatures of thermal events pertaining to TR phenomena

(Use of BTC HEL model 500)• First series of pouch cells tested in BTC, exploitation pending…

Fire/ TR issues (prediction/propagation/ignition) modeling making use of

various tools such as:

▪ TR multiphysics 0D/3D thermal runway model (coded with COMSOL®)(improved model from work of Sara Abada et al, )

(thermal abuse/Fire) Safety Modeling

❑ Other interactive modeling exercises with tools to obtain

guidance/response of design option at various level of integration

of building blocks of the EV (just started) up to recycling issues

▪ use of Simcenter Amsesim (IFP-EN), based on Siemens PLM

software

▪ Coupled modeling between INERIS and SAFT with Firefoam

v2.4 and NX Simcenter V 11.0.2 for TR propagation issue within

pack, with input data from INERIS and IFP-EN)

▪ (possibly) scenario based modeling of EV incident, such as fire-

induced toxicty in garage ?

• See Lecocq et al, J. of Power Source,

(Thermal abuse/fire) safety modeling (Cont’d)

Fire behavior comparison in an open field, unfortunately

• Battery tests are not designed for CFD code validation

The flame is 3 times longer than the cell and as large as the module

• Flame length is quite correct

• Temperature is in the correct order of magnitude

• But how is representative the boundary condition?

Orientation simulation CFD test at INERIS

The cell opening mode is crucial

• The whole back face

• A part of the back face

• An opening along the external surface

But inside a module the flame geometry is complex;;;

22 cm

5.4 cm

❑ The DEMOBASE project offer a unique experience /partnership to promote

safer, faster to market of innovative affordable Evs

▪ looping process

▪ Making use of interactive testing and modeling

❑ Paperwork also welcomed by partners in terms of guidance from the

safety viewpoint, as well as revised APR▪ Some safety goals appear challenging, such as fail-safe module or pack in case of

TR in one cell

▪ Post-crash scenario deserve further examination and show importance of pack

localisation and mechanical protection againt impacts and deformation during car

crash

▪ However, EV car will happen, as it happens on a regular basis with ICE cars, incl.

most exotic ones ! IN addition, there is no evidence of increased frequency of fire

event in existing EV fleet by comparison of ICE fleet,

❑ A reminder:▪ still many results to come, last but not least, check of performance of genuine

concept car issuing consolidated design process developed inside the consortium

Conclusions / perspectives

❑ Guy.marlair@ineris.fr pietro.perlo@ifevs.com

❑ Amandine.lecocq@ineris.fr philippe.desprez@saftbatteries.com

❑ Benjamin.truchot@ineris.fr

❑ Martin.petit@ifpen.fr

Thank you

mailto:Guy.marlair@ineris.frmailto:pietro.perlo@ifevs.commailto:Amandine.lecocq@ineris.frmailto:philippe.desprez@saftbatteries.commailto:Benjamin.truchot@ineris.frmailto:Martin.petit@ifpen.fr